Summary of Learning to Generate Conditional Tri-plane For 3d-aware Expression Controllable Portrait Animation, by Taekyung Ki et al.
Learning to Generate Conditional Tri-plane for 3D-aware Expression Controllable Portrait Animation
by Taekyung Ki, Dongchan Min, Gyeongsu Chae
First submitted to arxiv on: 31 Mar 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Multimedia (cs.MM)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel one-shot method for 3D-aware portrait animation is introduced, enabling control over facial expressions and camera views. The Export3D approach utilizes a tri-plane generator with an expression conditioning mechanism to generate 3D priors from given portrait images. This contrasts existing methods that rely on image warping, making it challenging to disentangle appearance and expression. A contrastive pre-training framework is proposed to eliminate unwanted appearance swaps when transferring expressions across identities. Experimental results demonstrate the effectiveness of this approach in generating controllable 3D-aware portrait images without appearance swap. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to create animated portraits from photos has been developed. This method, called Export3D, lets you control how a person’s face looks and what camera angle is used. It does this by creating a special kind of picture that shows the 3D shape of the face. To make sure this works well for different people, the authors came up with a way to separate a person’s appearance from their expressions, so that you can change the expression without changing the rest of the face. |
Keywords
» Artificial intelligence » One shot
